Acta Ophthalmologica 2013

Rates of visual field progression

in clinical glaucoma care
Anders Heijl,1 Patricia Buchholz,2 Gunilla Norrgren3 and
Boel Bengtsson1
1
Department of Clinical Sciences Malmö, Ophthalmology, Skåne University
Hospital, Lund University, Malmö, Sweden
2
Department for Medical and Pharmaceutical Issues, Regional Council Karlsruhe,
Karlsruhe, Germany
3
Allergan Ltd., Marlow International, The Parkway, Marlow, Buckinghamshire, UK

ABSTRACT. et al. 2003; Zahari et al. 2006; Leske

Purpose: To investigate rates of visual field progression and factors associated et al. 2007). Detailed data on rates of
with progression rate in open-angle glaucoma in clinical glaucoma care. progression have been reported for
Methods: We performed a retrospective chart review of all patients with mani- untreated glaucoma, that is, the
fest primary open-angle glaucoma (POAG) and pseudoexfoliation glaucoma natural history of progression of glau-
(PEXG) followed ‡ 5 years with ‡5 SITA Standard fields. Exclusion criteria comatous visual field loss (VFL) in
were minimal. Demographics, intraocular pressure values (IOP), treatment the prospective trials Collaborative
Normal-Tension Glaucoma Study
and treatment changes, and visual field (VF) data were recorded. VF progres-
(CNTGS 1998) and Early Manifest
sion rates were calculated as slopes of mean deviation (MD) over time.
Glaucoma Trial (EMGT; Leske et al.
Results: Five hundred and eighty-three patients were eligible. Three hundred 1999; Anderson et al. 2001; Heijl et al.
and sixty-seven (62%) had POAG and 221 (38%) PEXG. Median MD at 2009). Data on rates of progression in
study start was )10.0 dB. Mean follow-up time was 7.8 years (SD ± 1.2); ordinary clinical glaucoma care have
mean number of VF tests was 8.9 (SD ± 2.8). Progression rates varied very been sparse until recently when impor-
much among patients with a mean of )0.80 dB ⁄ year (SD ± 0.82; median tant results have been reported from
rate, )0.62), and 5.6% of patients progressed at rates worse than )2.5 dB per New York (Ahrlich et al. 2010; De
year A negative slope of MD values was observed in 89% of patients. Mean Moraes et al. 2011; Forchheimer et al.
IOP of all visits decreased over the study period from 20.15 to 18.10 mmHg. 2011).
Higher age and mean IOP, and more intensive treatment were associated with Rate of disease progression is one of
more rapid progression, while PEXG and IOP variation were not, if treatment the most important factors determining
intensity was taken into account. the risk of visual disability or blindness
Conclusion: Rates of visual field progression in manifest glaucoma with field in glaucoma and several guidelines for
loss in ordinary clinical care were highly variable. Progression rates rapid glaucoma management recommend
enough to influence quality of life were common. assessment of rate of progression in
routine glaucoma care (European
Glaucoma Society 2008; Heijl et al.
Key words: Glaucoma – progression – rate of progression – visual field
2010). Particularly against that back-
ground, it is desirable to learn more
Acta Ophthalmol. 2013: 91: 406–412 about rates of progression under ordin-
ª 2012 The Authors
Acta Ophthalmologica ª 2012 Acta Ophthalmologica Scandinavica Foundation
ary clinical care and of risk factors
associated with rate of progression.
doi: 10.1111/j.1755-3768.2012.02492.x In Sweden, most glaucoma care is
Re-use of this article is permitted in accordance with the Terms and Conditions set out at delivered in the public sector, which is
http://wileyonlinelibrary.com/onlineopen#OnlineOpen_Terms. an unusual feature of eye care in other
countries. As a matter of fact, most
ods visual field defects progress in a primary glaucoma care in our catch-
Introduction large proportion of glaucoma patients ment area is delivered at our Depart-
The large prospective clinical glau- and that progression rates vary ment of Ophthalmology at Skåne
coma trials and other studies have very much among patients (Gliklich University Hospital in Malmö. We
demonstrated that over longtime peri- et al. 1989; Membrey et al. 2000; Eid therefore considered it meaningful to

406

study glaucoma patients who receive
glaucoma care at our department,
because our glaucoma patients do not
represent a selection of patients with
particular characteristics, for example,
patients with particularly aggressive or
advanced disease.
Thus, the aim of this study was to
study perimetric rates of progression
and factors associated with progres-
sion in ordinary patients with open-
angle glaucoma.
Methods
The study was a retrospective review
of patient charts.
Inclusion and exclusion criteria, extracted
data
We studied records of patients with
diagnoses of primary open-angle
glaucoma (POAG) or pseudoexfolia-
tion (PEX) glaucoma. Potential
patients were identified with the help
of the computerized patient booking
system of our hospital (Skåne Uni-
versity Hospital, Malmö, Sweden),
comprising all visits during the last
decades. Diagnoses are registered for
all patients and visits, also for outpa-
tient visits. The hospital provides pri-
mary glaucoma care for a majority
of glaucoma patients (approximately
¾ of diagnosed patients) in the
catchment area (population 300 000)
in Southern Sweden. The study was
performed in accordance with the
declaration of Helsinki, and approval
was obtained from the Ethical
Review Board at Lund University,
Sweden. Advertisements were pub-
lished in local newspapers to allow
glaucoma patients who had visited
the department not to have data
from their patient records included in
the study.
Patient records were extracted for
patients who had been followed for at
least 5 years during the study period,
from March 1996 to August 2005.
The study period was selected in this
way because the SITA Standard (Ben-
gtsson et al. 1997) was introduced as
the standard perimetric test in our
department from March 1996, while
data retrieval started in September
2005, and because patient records on
paper were available for the whole
study period.
In this study, glaucoma diagnosis
required the presence of repeatable
visual field defects in SITA Standard
fields as defined by Glaucoma Hemi-
field Test results ‘Outside Normal
Limits’. The diagnosis POAG also
required open chamber angles,
absence of exfoliation or pigment dis-
persion syndrome or signs of other
secondary glaucomas. A glaucoma-
tous eye was classified as having
pseudoexfoliation glaucoma (PEXG)
if there was any report of PEX syn-
drome in that eye at any visit.
Exclusion criteria were kept to a
minimum: patients with other oph-
thalmic co-morbidity except cataract
with serious influence on visual field
results were not eligible. Other exclu-
sion criteria were blindness in the
study eye at study start, participation
in the EMGT (n = 180; Leske et al.
1999) or that the patient did not want
to contribute with data in the chart
review (n = 1).
Eligibility required continuous fol-
low-up for ‡5 years and that ‡5 Hum-
phrey SITA Standard visual field tests
were available.
One study eye was identified for
each patient. The study eye was the
eye with VFL or, if both eyes had
VFL, the eye with the largest VFL
defined by the global perimetric mean
deviation (MD) index.
Clinical parameters collected from
patient charts and reported in this
paper were age; gender; and – from
study eyes – PEX status, MD values
for all perimetric tests performed dur-
ing the study period, medication and
changes in medication, incisional and
laser (ALT) surgeries, and concomi-
tant eye disease.
Statistical analysis
Descriptive analyses were performed
on demographics, follow-up time,
baseline MD and on intraocular pres-
sure (IOP), IOP variation defined as
the range between the highest and
lowest IOP values measured during
the study period, and MD values over
time. Study start for a patient was
defined as the first visit that contrib-
uted records data from the patient.
For patients in whom follow-up
started before 1996, the study start
was thus usually in 1996, that is, when
the first SITA Standard fields were
obtained. For patients diagnosed after
Acta Ophthalmologica 2013
1996, the baseline usually was the first
visit after diagnosis.
For each study eye, we calculated
visual field progression rate using lin-
ear regression analysis of perimetric
MD values over time, where rate of
progression is the slope expressed in
dB ⁄ year. Reliability criteria, that is,
fixation losses, false positive or false
negative answers, were not taken into
account, and only a few obviously
artefactual fields, for example, clover
leaf fields or similar extreme outliers,
were excluded from analysis. We also
calculated mean IOP over all study
visits and IOP variation.
Changes in drug treatment were reg-
istered and compounded into a drug
change score. Each change added one
to this score, so that a patient who
was on the same drug treatment dur-
ing the study period received a score
of 0, while a patient who was switched
to another drug or in whom one drug
was added, got a score of 1, and a
patient who encountered three changes
received a score of 3.
We studied factors that we assumed
might be associated with progression
rate. In a first multiple linear regres-
sion analysis, we included age, gender,
mean IOP, IOP variation, presence of
exfoliation syndrome (PEX) and MD
at study start. All parameters that
were not normally distributed (IOP
range, MD, drug change score) were
then divided by median split.
As treatment may change IOP range,
we subsequently performed a second
multivariate analysis adding treatment
variables: the drug treatment score, in-
cisional surgery and ALT.
In a third multivariate analysis, we
removed all variables that did not
reach the p < 0.10 significance level
in the second analysis.
All statistical analyses were per-
formed using spss v. 19.0 (IBM SPSS,
Chicago, IL, USA).
Results
All eligible patients (n = 583) were
included in the analysis. Mean age at
start was 71.4 years (min 31; max 95).
The majority of patients, 367 (63%),
were female.
Three hundred and sixty-seven eyes
(62%) had POAG, while 221 (38%)
had PEX glaucoma. The range of dis-
ease severity was large (Fig. 1).
Expressed as MD values in the study
407
Acta Ophthalmologica 2013

70 negative; 60% statistically significant start of the follow-up period was asso-
at the p < 0.05 level. ciated with slower progression, possi-
60 bly because of truncation (cf.
50
Discussion).
Intraocular pressure
Number of eyes

When the treatment parameters

40
30
20
10
0 IOP variation (highest – lowest
–32 –28 –24 –20 –16 –12
MD (dB)
Fig. 1. Mean deviation (MD) at the begin-
ning of the follow-up period. The whole spec-
trum of glaucomatous field loss is covered.
The median MD ()10.0 dB) corresponds to
moderately advanced glaucoma.
eye at study start, severity ranged
from )30.4 to +1.6 dB; the median
MD corresponded to moderately
advanced glaucoma, )10.0 dB (Mills
et al. 2006).
Mean follow-up time for the study
Mean IOP and SD for each study
year are shown in Fig. 3 for all eyes
and for eyes with POAG and PEX
glaucoma. Mean IOP in the whole
cohort decreased from 20.15 mmHg in
1996 to 18.10 mmHg in 2005 (Fig. 3).
–8 –4 0 recorded IOP) varied among study
eyes from 2 to 61 mmHg, with a med-
ian of 13.
Medications and surgery
The median number of additions and
changes in medication was 3; mini-
mum 0 and maximum 30. Argon laser
trabeculoplasty was performed in 179
eyes, once in 127 eyes, twice in 48 and
three times in four eyes. Sixty eyes
underwent incisional surgeries (all tra-
beculectomies), all once only.
medications, ALT and trabeculecto-
mies were added to the multivariate
analysis, IOP range was no longer sta-
tistically significant (p = 0.101) while
mean IOP (p = 0.033), age
(p = 0.000) and MD remained signifi-
cant. As expected, this analysis also
showed that changes in medication,
ALTs and trabeculectomies were more
common in eyes with steeper negative
slopes (drug change score p = 0.003;
ALT p = 0.041; trabeculectomies
p = 0.040).
The results of the third multivariate
analysis including only factors signifi-
cant at the p < 0.1 level in the second
30 All
25

Mean IOP (mmHg)

20
period was 7.8 years (SD ± 1.2;
range, 5.0–9.6 years). The mean num- 15
Factors associated with progression
ber of visual field tests per eye was 8.9
10
(SD ± 2.8; range, 5–25). The results of the first multiple linear
Progression rates are clear from regression analysis are shown in 5
Fig. 2. The mean MD slope was Table 1. Older age, higher mean IOP
0
)0.80 dB ⁄ year (SD ± 0.82, median, and IOP range were all associated (A) 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Year
)0.62), and slopes ranged from )5.58 with significantly faster progression
to +1.24 dB ⁄ year (Fig. 3) with a (more negative slopes), while gender
30 POAG
negatively skewed distribution. and presence of pseudoexfoliation
Eighty-nine percentage of slopes were syndrome were not. More field loss at 25
Mean IOP (mmHg)

20

90 15

10
80
5
70
(B) 0
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Year
60

30 PEX
Frequency

50
25
Mean IOP (mmHg)

40 20

15
30
10
20
5

10 (C) 0
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Year
0
–6.0 –5.6 –5.2 –4.8 –4.4 –4.0 –3.6 –3.2 –2.8 –2.4 –2.0 –1.6 –1.2 –0.8 –0.4 0 +0.4 +0.8 +1.2 Fig. 3. Mean intraocular pressure (IOP) over
Rate of progression (dB/year) time ±1 standard deviation for all eyes (A),
for the group of eyes with primary open-
Fig. 2. Rates of progression expressed in dB ⁄ year. The distribution is wide and has a negative angle glaucoma (B) and exfoliation glaucoma
tail of rapidly progressing eyes. (C). IOP decreased slowly over time.

408

Table 1. Results of multivariate analysis of factors associated with rate of progression not
including treatment parameters.
Variable Reference
Age N⁄A
Mean IOP N⁄A
IOP range* <12 mmHg
MD* Worse than )10.03 dB
PEX syndrome No
Gender Male
IOP, intraocular pressure; MD, mean deviation.
* Divided by median split.
Yes ⁄ no.
Table 2. Results of multivariate analysis of factors associated with rate of progression when
including treatment parameters.
Variable Reference
Age N⁄A
Mean IOP N⁄A
MD* Worse than )10.03 dB
ALT No
Trabeculectomies No
IOP, intraocular pressure; MD, mean deviation.
* Divided by median split.
Yes ⁄ no.
multivariate analysis are shown in
Table 2. Age, mean IOP, MD at study
start, ALT and trabeculectomies were
all significantly associated with more
rapid progression.
Discussion
In this large group of patients treated
for open-angle glaucoma, a majority
progressed during a mean follow-up
time of 7.8 years.
Rate of progression varied very
much among patients; the median rate
of progression corresponded to going
from a normal to a blind field in
approximately 50 years, but a consid-
erable minority progressed much fas-
ter. Fast progression, for example
>1.0 dB ⁄ year, was relatively common
(Fig. 2). Progression at such rates is
clearly important for quality of life. In
10–15 years, eyes progressing at such
rapid rates would go from early glau-
coma to advanced glaucoma or from
moderate to severe disease (Mills et al.
2006).
Study results confirmed mean IOP
as risk factors for glaucoma progres-
sion. In initial analyses, IOP range
was also a significant risk factor in
first multivariate analyses, but when
factors indicating treatment and treat-
ment changes were added to the anal-
Slope Significance
)0.019 0.000
)0.036 0.001
)0.256 0.000
)0.175 0.009
)0.071 0.206
0.078 0.250
Slope Significance
)0.021 0.000
)0.028 0.011
)0.188 0.004
)0.182 0.021
)0.283 0.012
yses, the IOP range was no longer
significant.
This study has strengths and weak-
nesses. Among the strengths is the
large number of patients analysed.
The data are also quite representative
for all treated glaucoma patients in
the catchment area. This is attribut-
able to the unusual structure of Swed-
ish ophthalmic care, where a single
university hospital site can provide
most primary glaucoma care in a large
catchment area. In most other EU
countries, this type of data is not
obtainable as glaucoma patients have
access to multiple health care provid-
ers and receive treatment in a variety
of settings: office-based private prac-
tices, private clinics or state-owned
clinics ⁄ hospitals. Another strength is
that most patients came to regularly
scheduled appointments, which led to
a relatively large number of fields col-
lected over a time interval long
enough for reasonable assessments of
rate of progression. The fields were
also always obtained with the same
SITA Standard 30-2 test, eliminating
errors caused by different threshold
characteristics among tests (Heijl et al.
2000).
This is a retrospective study. This is
a weakness because much data, for
example, on systemic blood pressure
and cardiac disease are missing, which
Acta Ophthalmologica 2013
makes it impossible to analyse
whether such systemic health parame-
ters were associated with glaucoma
progression rate. At the same time,
the retrospective study design can be a
considered strength, resulting in data
that represent ordinary routine clinical
care. We consider it likely that with a
prospective study design, eyes with
higher progression rates would have
received more drastic changes of treat-
ment, including surgery, that might
have resulted in lower IOP values and
somewhat lower progression rates and
that adherence might have been higher
than in ordinary glaucoma care.
Treatment intensity and target pres-
sures may of course differ between
centres and will influence rate of pro-
gression. Even small differences in
IOP may make a difference. Thus,
several of the large prospective trials
of glaucoma and ocular hypertension
have shown that the risk of progres-
sion may on the average decrease by
10% or more for each mmHg of IOP
reduction (Gordon et al. 2002; Leske
et al. 2003; Miglior et al. 2007; Chau-
han et al. 2008). Even if percentage
reduction of risk is not identical to
reduction of rate of progression, they
are closely related. The Malmö
department did not have a strict, writ-
ten care programme for glaucoma,
and over 20 physicians were involved
in the glaucoma care. Experience var-
ied among physicians most of whom
were not glaucoma specialists.
IOP levels on treatment reflect the
level of treatment intensity. It is there-
fore of interest to compare the IOP
levels in the current study with other
published figures. In the current study,
mean IOP was just over 20 mmHg in
1996 and decreased by 2 mmHg dur-
ing the study period. These IOP val-
ues are similar to those found in other
chart analyses performed in Europe
during the same time period. In a con-
secutively recruited retrospective chart
analysis in Sweden and France of sev-
eral hundred patients with open-angle
glaucoma and ocular hypertension,
mean IOP over time in treated
patients was 18.2 ± 4.2 mmHg,
18.9 ± 4.9 in the Swedish group
(Lindblom et al. 2006). This is very
similar to the results of the current
results, particularly because our means
include some untreated IOP values in
newly diagnosed patients. A retrospec-
tive medical record analysis in several
409
Acta Ophthalmologica 2013
academic and office-based centres in
Sweden and the United States had
had very similar IOP levels at study
end, in USA 18.4 ± 4.3 mmHg; Swe-
den 18.8 ± 5.3 mmHg (Kobelt-
Nguyen et al. 1998). In a Dutch study
of 500 representative patients with
glaucoma and ocular hypertension,
the mean (±SD) IOP in the 355
patients with glaucoma was
20.9 ± 6.9 mmHg (Oostenbrink et al.
2001). Thus, IOP levels seem not to
have differed much among the current
and other published studies.
There are some recent comparable
studies that report rates of visual field
progression, particularly in clinical
settings. Several of those come from
the same research group in New
York. The closest comparison may be
a study by De Moraes et al. (2011).
This paper focuses on risk factors for
progression but also reports global
rates visual field progression expressed
as MD loss ⁄ year. The study has sev-
eral similarities with ours: it is of very
similar size, duration, and also num-
ber of fields, and the study population
is mainly white, with a majority of
women. Their patient group was quite
different, however. The authors
emphasize that they represent a ter-
tiary referral centre. They included a
more mixed patient group including
narrow angle and juvenile glaucoma,
and the patients were on the average
6.5 years younger at baseline, and
mean MD values were better.
As it is known that the risk of pro-
gression in many trials decreases by
10–19% per mmHg of IOP reduction
(Gordon et al. 2002; Leske et al. 2003;
Miglior et al. 2007; Chauhan et al.
2008), it might be permissible to
assume that rates of progression are
similarly influenced by IOP, particu-
larly because risk calculations in trials
often are based on analyses that are
based on time to progression. If mean
IOP in our cohort had been at the
same level as that in de Moraes’
group, and assuming that the risk ⁄ rate
reduction per mmHg is similar to that
discussed above, it is reasonable to
assume that our progression rates had
been 30–40% slower and thus quite
similar to those reported by Moraes.
We therefore feel that the results of
the two studies are in agreement. The
effect of age on progression has been
estimated, for example in a report
from the Canadian glaucoma study,
410
where the hazard ratio was 1.04 per
year of increasing age (Chauhan et al.
2010, 2008). Here, it was 1.09 per
year. A study by Forchheimer et al.
(2011) reports progression rates that
are similar to those of Moraes in a
similar group of patients. The aim of
the Forchheimer’s study was to study
the influence of baseline perimetric
status, and it is likely that there is
considerable overlap in the patient
groups studied by de Moraes and
Forchheimer.
We chose to include the worse eye
in patients with bilateral glaucoma,
mostly to be sure that study eyes
really had manifest glaucoma. This
may have increased the risk of trunca-
tion (floor effects), when visual fields
defects become more and more
advanced. All studies like the present
one have some problems with floor
effects, and, therefore, reported mean
rates of progression should be
regarded as minimum estimates.
The Canadian Glaucoma Study also
reported rates of progression (Chau-
han et al. 2010). Those rates were
much lower than ours, with a mean
rate in 45 progressing patients of
)0.35 dB ⁄ year and a slightly positive
change (+0.05 dB ⁄ year) in 153 nonp-
rogressing patients. Our results are
not at all in line with these Canadian
results, but the differences between
the studies were large. Thus, the
Canadian patients were approximately
8 years younger than the Swedish
patients, their mean IOP was much
lower at 14.8 mmHg, the proportion
of PEXG was also much lower, and
the patients had considerably earlier
glaucoma (study eligibility required
MD values better than )10 dB).
Another factor that may have contrib-
uted to the considerably better pro-
gression rates in the Canadian study
was that these patients were followed
in a well-organized, prospective study.
Patients in such studies are generally
assumed to have considerably better
adherence to prescribed therapy than
patients in routine medical care. It is
tempting to guess that this might be
the main factor, because mean age
and mean IOP are in fact quite similar
in the Canadian study and the studies
from New York (De Moraes et al.
2011; Forchheimer et al. 2011), while
progression rates are very different.
It is of some interest to also com-
pare our observed progression rates,
with such rates in untreated glau-
coma. Two studies provide such nat-
ural history progression rates, the
CNTGS and the EMGT (Anderson
et al. 2001; Heijl et al. 2009).
Reported mean rates in normal-ten-
sion glaucoma eyes are similar in
CNTGS and EMGT, approximately
0.4 dB ⁄ year. This is less than the
median rate in the current study of
treated eyes. Untreated rates differ
much between groups of glaucoma
patients, however. The mean rate in
EMGT was 1.08 dB ⁄ year, in high
tension glaucoma it was 1.31 dB ⁄ year
and in PEX glaucoma 3.13 dB ⁄ year.
It is therefore clear that despite the
fact that mean and median progres-
sion rates in the current study were
not slow, they were considerably
slower than in untreated patients
from the same population.
Reflecting on the generalizability of
the results, we must notice that all
data were collected at a single site
with mostly Caucasian patients and a
rather high percentage of exfoliation
glaucoma. We still believe that the
patient population is reasonably rep-
resentative of glaucoma patients in
this part of Europe. PEXG is com-
mon in Malmö, but also in the neigh-
bouring Nordic countries as in many
other parts of the world, for example,
Greece, Russia, Turkey, India, parts
of Africa (Ritch 2001; Ritch & Schlot-
zer-Schrehardt 2001). Stage of disease
could be of importance. In Sweden,
asymptomatic individuals only rarely
see ophthalmologists for check-ups;
individuals with a positive family his-
tory of glaucoma may be an excep-
tion. Glasses are often dispensed by
opticians with limited capability of
detecting glaucoma by means other
than with tonometry, which is often
performed by opticians in customers
over the age of 50. As a result, clinical
diagnoses of glaucoma are often made
late (Grødum et al. 2002). Mean MD
at study start was also worse
()10.0 dB) in this study than in sev-
eral other large studies reporting rates
of progression (Ahrlich et al. 2010;
Chauhan et al. 2010; De Moraes et al.
2011; Forchheimer et al. 2011).
Our analyses of factors associated
with progression confirm some other
risk factors that have been identified
in controlled trials. IOP was found to
be a risk factor in the EMGT (Leske
et al. 2003, 2007), AGIS (The AGIS

Investigators 2000), CIGTS (Lichter
et al. 2001) and Canadian Glaucoma
Study (Chauhan et al. 2008), while
EMGT and AGIS (The AGIS Investi-
gators 2002) also found age to be sig-
nificant factors. Age was a significant
factor also in the Canadian Glaucoma
Study (Chauhan et al. 2008). EMGT
results additionally defined presence
of PEX as a risk factor (Leske et al.
2007).
Several treatment trials have indi-
cated higher risk of progression in
eyes with worse MD values (The
AGIS Investigators 2002, Leske et al.
2007). In the current study, worse
visual field status was not a risk factor
for more rapid progression, but the
opposite. Worse baseline field status
was associated with slower measured
progression. It is likely that this was
attributable to truncation; a visual
field with very advance cannot pro-
gress as much as a field with smaller
defects. A recent study by Forchhei-
mer et al. (2011) found no difference
in progression rates depending on
MD after correcting for IOP.
The appearance and disappearance
of IOP range as a risk factor is very
interesting, however, considering the
widely divided opinions on this mat-
ter (Singh & Shrivastava 2009). In
our initial analyses, which did not
take treatment changes into account,
larger IOP range was a significant
risk factor, but when factors flagging
treatment intensity were included in
the analysis, the significance of range
disappeared. Instead, the results
showed that treatment intensity was
indeed positively correlated with
worse progression rates. The signifi-
cance when treatment changes were
unaccounted for is in agreement with
other studies, when results have been
analysed without correcting for such
changes (Nouri-Mahdavi et al. 2004;
Bengtsson et al. 2007; Singh & Shriv-
astava 2009). Also, the AGIS investi-
gators found that IOP variation was
less important and only significant in
eyes with low IOP levels, when they
limited their analysis to eyes that
had only one intervention (Caprioli
& Coleman 2008). We have previ-
ously shown that if treatment
changes are accounted for apparent
significance of IOP variations may
disappear, while mean IOP remained
a highly significant factor (Bengtsson
et al. 2007).
Our drug change score can be criti-
cized for being arbitrary and not dif-
ferentiating between changes initiated
to further decrease IOP or for other
reasons, for example, encountered side
effects. In a retrospective chart analy-
sis like the present one, it is not
always clear why a prescription has
been changed, however, but the
change itself is clearly documented.
This is one reason why we decided to
simply add management changes.
Another advantage is that this
approach leaves no room for subjec-
tive biases when data are extracted
and analysed.
The results support the notion that
this study like any other study that
allows treatment changes is not really
suited to investigate the effect IOP
variation. If clinical care is delivered
in an optimal way, one would expect
treating physicians to intensify treat-
ment in progressing patients, thus pro-
ducing a larger IOP variation in eyes
with more rapid progression. Our
analyses showed that this was indeed
the case.
Thus, we found that rates of visual
field progression in manifest glaucoma
with field loss in ordinary clinical care
were highly variable among patients.
Most study eyes progressed, and pro-
gression rates that were rapid enough
to influence quality of life over a 10–
15 year period were common. The risk
factors for progression were higher
mean IOP and older age. Presence of
pseudoexfoliations was not a risk fac-
tor in a multivariate analysis, nor was
IOP range, if treatment factors were
included in the analyses.
Acknowledgements
This study was supported by the
Swedish Research Council K2005-
74X-10426-13A, by the Järnhardt
Foundation and by an unrestricted
grant from Allergan Inc., Irvine CA,
USA. This study was presented in
part at ARVO Fort Lauderdale, 28
April 2008.
References
Ahrlich KD, de Moraes CG, Teng CC, Prata
TS, Tello C, Ritch R & Liebmann JM
(2010): Visual field progression differences
between normal-tension and exfoliative
high-tension glaucoma. Invest Ophthalmol
Vis Sci 51: 1458–1463.
Acta Ophthalmologica 2013
Anderson DR, Drance SM & Schulzer M
(2001): Natural history of normal-tension
glaucoma. Ophthalmology 108: 247–253.
Bengtsson B, Olsson J, Heijl A & Rootzen H
(1997): A new generation of algorithms for
computerized threshold perimetry, SITA.
Acta Ophthalmol Scand 75: 368–375.
Bengtsson B, Leske MC, Hyman L & Heijl A
(2007): Fluctuation of intraocular pressure
and glaucoma progression in the early
manifest glaucoma trial. Ophthalmology
114: 205–209.
Caprioli J & Coleman AL (2008): Intraocular
pressure fluctuation a risk factor for visual
field progression at low intraocular pres-
sures in the advanced glaucoma interven-
tion study. Ophthalmology 115: 1123–1129
e1123.
Chauhan BC, Mikelberg FS, Balaszi AG,
LeBlanc RP, Lesk MR & Trope GE (2008):
Canadian Glaucoma Study: 2. risk factors
for the progression of open-angle glaucoma.
Arch Ophthalmol 126: 1030–1036.
Chauhan BC, Mikelberg FS, Artes PH, Bala-
zsi AG, LeBlanc RP, Lesk MR, Nicolela
MT & Trope GE (2010): Canadian glau-
coma study: 3. Impact of risk factors and
intraocular pressure reduction on the rates
of visual field change. Arch Ophthalmol
128: 1249–1255.
Collaborative Normal-Tension Glaucoma
Study Group (1998): The effectiveness of
intraocular pressure reduction in the treat-
ment of normal-tension glaucoma. Am J
Ophthalmol 126: 498–505.
De Moraes CG, Juthani VJ, Liebmann JM,
Teng CC, Tello C, Susanna R & Ritch R
(2011): Risk factors for visual field progres-
sion in treated glaucoma patients. Arch
Ophthalmol 129: 562–568.
Eid TM, Spaeth GL, Bitterman A & Stein-
mann WC (2003): Rate and amount of
visual loss in 102 patients with open-angle
glaucoma followed up for at least 15 years.
Ophthalmology 110: 900–907.
European Glaucoma Society (2008): Termi-
nology and guidelines for glaucoma.
Savona: Editrice Dogam S.r.l.
Forchheimer I, de Moraes CG, Teng CC,
Folgar F, Tello C, Ritch R & Liebmann
JM (2011): Baseline mean deviation and
rates of visual field change in treated glau-
coma patients. Eye 25: 626–632.
Gliklich RE, Steinmann WC & Spaeth GL
(1989): Visual field change in low-tension
glaucoma over a five-year follow-up. Oph-
thalmology 96: 316–320.
Gordon MO, Beiser JA, Brandt JD et al.
(2002): The Ocular Hypertension Treat-
ment Study: baseline factors that predict
the onset of primary open-angle glaucoma.
Arch Ophthalmol 120: 714–720; discussion
829–730.
Grødum K, Heijl A & Bengtsson B (2002): A
comparison of glaucoma patients identified
through mass screening and in routine clin-
ical practice. Acta Ophthalmol Scand 805:
627–631.
Heijl A, Bengtsson B & Patella VM (2000):
Glaucoma follow-up when converting from
411
Acta Ophthalmologica 2013
long to short perimetric threshold tests.
Arch Ophthalmol 118: 489–493.
Heijl A, Bengtsson B, Hyman L & Leske MC
(2009): Natural history of open-angle glau-
coma. Ophthalmology 116: 2271–2276.
Heijl A, Alm A, Bengtsson B, Bergström A,
Calissendorff B, Lindblom B & Lindén C
(2010): Riktlinjer för glaukomsjukvården.
Malmö: Sveriges Ögonläkarförening,
Bäcklunds Media.
Kobelt-Nguyen G, Gerdtham UG & Alm A
(1998): Costs of treating primary open-
angle glaucoma and ocular hypertension: a
retrospective, observational two-year chart
review of newly diagnosed patients in Swe-
den and the United States. J Glaucoma 7:
95–104.
Leske MC, Heijl A, Hyman L & Bengtsson B
(1999): Early Manifest Glaucoma Trial:
design and baseline data. Ophthalmology
106: 2144–2153.
Leske MC, Heijl A, Hussein M, Bengtsson B,
Hyman L & Komaroff E (2003): Factors
for glaucoma progression and the effect of
treatment: the early manifest glaucoma
trial. Arch Ophthalmol 121: 48–56.
Leske MC, Heijl A, Hyman L, Bengtsson B,
Dong L & Yang Z (2007): Predictors of
long-term progression in the early manifest
glaucoma trial. Ophthalmology 114: 1965–
1972.
Lichter PR, Musch DC, Gillespie BW, Guire
KE, Janz NK, Wren PA & Mills RP
(2001): Interim clinical outcomes in the
Collaborative Initial Glaucoma Treatment
Study comparing initial treatment random-
ized to medications or surgery. Ophthal-
mology 108: 1943–1953.
412
Lindblom B, Nordmann JP, Sellem E et al.
(2006): A multicentre, retrospective study
of resource utilization and costs associated
with glaucoma management in France and
Sweden. Acta Ophthalmol Scand 84: 74–
83.
Membrey WL, Poinoosawmy DP, Bunce C,
Fitzke FW & Hitchings RA (2000): Com-
parison of visual field progression in
patients with normal pressure glaucoma
between eyes with and without visual field
loss that threatens fixation. Br J Ophthal-
mol 84: 1154–1158.
Miglior S, Pfeiffer N, Torri V, Zeyen T,
Cunha-Vaz J & Adamsons I (2007): Predic-
tive factors for open-angle glaucoma
among patients with ocular hypertension in
the European Glaucoma Prevention Study.
Ophthalmology 114: 3–9.
Mills RP, Budenz DL, Lee PP, Noecker RJ,
Walt JG, Siegartel LR, Evans SJ & Doyle
JJ (2006): Categorizing the stage of glau-
coma from pre-diagnosis to end-stage dis-
ease. Am J Ophthalmol 141: 24–30.
Nouri-Mahdavi K, Hoffman D, Coleman
AL, Liu G, Li G, Gaasterland D & Capri-
oli J (2004): Predictive factors for glauco-
matous visual field progression in the
Advanced Glaucoma Intervention Study.
Ophthalmology 111: 1627–1635.
Oostenbrink JB, Rutten-van Molken MP &
Sluyter-Opdenoordt TS (2001): Resource
use and costs of patients with glaucoma or
ocular hypertension: a one-year study
based on retrospective chart review in the
Netherlands. J Glaucoma 10: 184–191.
Ritch R (2001): Exfoliation syndrome. Curr
Opin Ophthalmol 12: 124–130.
Ritch R & Schlotzer-Schrehardt U (2001):
Exfoliation syndrome. Surv Ophthalmol
45: 265–315.
Singh K & Shrivastava A (2009): Intraocular
pressure fluctuations: how much do they
matter? Curr Opin Ophthalmol 20: 84–87.
The AGIS Investigators (2000): The
Advanced Glaucoma Intervention Study
(AGIS): 7. The relationship between control
of intraocular pressure and visual field dete-
rioration. Am J Ophthalmol 130: 429–440.
The AGIS Investigators (2002): The
Advanced Glaucoma Intervention Study
(AGIS): 12. Baseline risk factors for sus-
tained loss of visual field and visual acuity
in patients with advanced glaucoma. Am J
Ophthalmol 134: 499–512.
Zahari M, Mukesh BN, Rait JL, Taylor HR
& McCarty CA (2006): Progression of
visual field loss in open angle glaucoma in
the Melbourne Visual Impairment Project.
Clin Experiment Ophthalmol 34: 20–26.
Received on February 29th, 2012.
Accepted on May 30th 2012.
Correspondence:
Anders Heijl
Department of Ophthalmology
Skåne University Hospital
Lund University
SE-20502 Malmö
Sweden
Tel: + 46 40 332741
Fax: + 46 40 336212
Email: anders.heijl@med.lu.se